Here is a snapshot 📸 of your model outputs for run ID 20250310_220042EDT, from config config_sample_2pop_inference.yml, stored in model_output.

1 Infection timeseries: SEIR model output

These are the outputs for the compartmental epidemic model, stored in the seir directory, which track the prevalence and incidence of individuals in each model compartment over time.

Incidence values are per day.

## [1] "Assuming inference run with files in/global/final"
## [1] "Importing seir files (n = 5):"
## [1] "Assuming inference run with files in/global/final"
## [1] "Importing llik files (n = 5):"

1.1 All infections

Total number of individuals in each infection state over time (compartments defined by infection_stage), aggregated across other strata. Plotted for slot 5 which has the highest total likelihood over all subpopulations (if inferrence was run) or was randomly chosen (if no inference).

1.1.1 Total

1.1.1.1 Prevalence

1.1.1.2 Cumulative incidence

1.1.2 Per capita

1.1.2.1 Prevalence

1.1.2.2 Cumulative incidence

2 Outcome timeseries: HOSP model output

These are the outputs for the observational (“outcomes”) model, stored in the hosp directory, which tracks the incidence and prevalence of individuals with defined observed disease outcomes over time.

2.1 Aggregate outcomes - by slot

Total number of individuals with each outcome over time, aggregated across other strata (only outcomes without an “_” specifying a stratification are plotted). If more than one simulation (slot) was run, results are plotted for slot 5 which has the highest total likelihood over all subpopulations (if inference was run) or was randomly chosen (if no inference). Incidence values are per day.

If inference was run, only some of these outcomes may have been used in inference, and the outcomes may have been aggregated to a longer time period (e.g., weeks, months). Inference-specific outcomes, along with the data they were compared to, are shown in later plots.

[1] “Assuming inference run with files in/global/final” [1] “Importing hosp files (n = 5):”

2.1.1 incidCase

2.1.1.1 Total

2.1.2 incidHosp

2.1.2.1 Total

2.1.3 currHosp

2.1.3.1 Total

2.1.4 incidDeath

2.1.4.1 Total

2.2 Inference-specific outcomes - by slot

The inference method specified that the model be fit to sum_hosp, with aggregation over period: 1 weeks. Plotted for slot 5 which has the highest total likelihood over all subpopulations (if inference was run) or was randomly chosen (if no inference).

[1] “Assuming inference run with files in/global/final” [1] “Importing hosp files (n = 5):”

2.2.1 sum_hosp

2.3 Inference-specific outcomes - quantiles

The inference method specified that the model be fit to sum_hosp, with aggregation over period: 1 weeks. In total 5 slots ran successfully.

2.3.1 sum_hosp

## Inference-specific outcomes - by likelihood{.tabset}

The inference method specified that the model be fit to sum_hosp, with aggregation over period: 1 weeks. In total 5 slots ran successfully.

This section plots the top 5 and bottom 5 log likelihoods for each subpopulation.

2.3.2 sum_hosp

3 Infection model parameters: SNPI model output

These are the parameters that define time-dependent modifications to the infection model parameters, and are stored in the snpi directory.

3.1 Values by slot

If inference is run, parameters are the final values at the end of all MCMC iterations, colored by their likelihoods in a given subpopulation.

3.2 MCMC evolution

The accepted value of the parameter for each iteration of the MCMC algorithm, colored by their likelihood in a given subpopulation. If more than 5 slots were run, we will plot only the top 5 and bottom 5 log likelihoods for each subpopulation.

3.3 MCMC evolution - chimeric vs global

The accepted value of the parameter for each iteration of the MCMC algorithm, for both the chimeric and global chain, in a given subpopulation. Plotted for slot 5 which has the highest total likelihood over all subpopulations (if inference was run) or was randomly chosen (if no inference).

4 Outcome model parameters: HNPI model output

This shows the parameters associated with your outcomes model, for all subpopulations.

4.1 Values by slot

If inference is run, parameters are the final values at the end of all MCMC iterations, coloured by their likelihoods in a given subpopulation.

4.2 MCMC evolution

The accepted value of the parameter for each iteration of the MCMC algorithm, colored by their likelihood in a given subpopulation. If more than 5 slots were run, we will plot only the top 5 and bottom 5 log likelihoods for each subpopulation.

4.3 MCMC evolution - chimeric vs global

The accepted value of the parameter for each iteration of the MCMC algorithm, for both the chimeric and global chain, in a given subpopulation. Plotted for slot 5 which has the highest total likelihood over all subpopulations (if inference was run) or was randomly chosen (if no inference).

5 Likelihood: LLIK model output

5.1 Acceptance and likelihood trajectories - All slots and subpopulations

This plot shows the binary acceptance decision for each MCMC iteration (accept), the probability of acceptance for that acceptance decision (accept_prob), the running average acceptance probability (accept_avg), and the likelihood. Chimeric values are subpopulation specific - there are likely more acceptances as well as acceptances that can increase subpop-specific likelihood while not changing the total likelihood. Global acceptances occur for all subpopulations together, and will always result in the total likelihood increasing, but could result in decreases in the subpop-specific likelihood.

## [1] "Assuming inference run with files in/global/intermediate"
## [1] "Importing llik files (n = 760):"
## [1] "Assuming inference run with files in/chimeric/intermediate"
## [1] "Importing llik files (n = 760):"

5.2 Acceptance and likelihood trajectories - Single slot